---
title: 'Sum Small: Medical Dialogue to SOAP Summarizer'
emoji: 📄
colorFrom: green
colorTo: pink
sdk: static
pinned: false
license: mit
datasets:
- omi-health/medical-dialogue-to-soap-summary
language:
- en
metrics:
- rouge
---

# Model Card for Sum (3B) Small                                                                                             

> **Research article:** [Omi-Sum 3B: Open Clinical SOAP Model →                                                             
omi.health/research/omi-sum](https://omi.health/research/omi-sum)
> Built by [Omi Health](https://omi.health) · [Scribe product page](https://omi.health/scribe) · [All                       
research](https://omi.health/research)                                                                                      

## Model Description                                                                                                        
                                                        
Sum Small is a language model specifically designed to generate SOAP summaries from medical dialogues. It is a fine-tuned   
version of [Microsoft/Phi-3-mini-4k-instruct](https://huggingface.co/microsoft/Phi-3-mini-4k-instruct), trained on the
[omi-health/medical-dialogue-to-soap-summary](https://huggingface.co/datasets/omi-health/medical-dialogue-to-soap-summary)  
dataset. It demonstrates competitive performance against larger closed models.

Full benchmark write-up, methodology, and reproducibility notes:                                                            
**[omi.health/research/omi-sum](https://omi.health/research/omi-sum)**.
                                                                                                                          
## Intended Use                                           

This model is intended for research and development in AI-powered medical documentation. It is **not** ready for direct     
clinical use without further validation and should be integrated with additional safety guardrails before deployment in a
medical setting.                                                                                                            
                                                        
## Training Data

The model was trained on [Omi Health's](https://omi.health) synthetic `medical-dialogue-to-soap-summary` dataset, which     
consists of 10,000 synthetically generated dialogues and corresponding SOAP summaries.
                                                                                                                          
## Training Procedure                                     

Training was conducted on NVIDIA A100 GPUs.                                                                                 

## Evaluation                                                                                                               
                                                        
The performance of Sum Small has been evaluated using ROUGE-1 as follows:                                                   

| Model                      | ROUGE-1 |                                                                                    
|----------------------------|---------|                  
| **Omi-Sum 3B Small**       | **70**  |
| GPT-4 Turbo                | 69      |
| Llama-3 8B Instruct        | 59      |                                                                                    
| Phi-3 3B mini 4k instruct  | 55      |
| GPT-3.5                    | 54      |                                                                                    
| Phi-2 basic                | 41      |                  
                                                                                                                          
Full results, evaluation methodology, and safety analysis:                                                                  
[omi.health/research/omi-sum](https://omi.health/research/omi-sum).
                                                                                                                          
## Limitations                                            

While Sum Small demonstrates promising results, the training data is synthetic and not derived from actual clinical         
interactions. Care must be taken when considering this model for practical applications, as it requires significant testing
and adaptation to meet clinical safety standards.                                                                           
                                                        
## Licensing

Released under the MIT License — free for commercial and non-commercial use.                                                

## Citation                                                                                                                 
                                                        
If you use this model or reference the benchmark, please cite:                                                              

**APA** — Omi Health. (2024). *Omi-Sum 3B: Open-Source Model for Medical Summaries*. https://omi.health/research/omi-sum    
                                                        
**BibTeX**                                                                                                                  
```bibtex                                                 
@misc{omi_sum_3b_2024,
title   = {Omi-Sum 3B: Open-Source Model for Medical Summaries},                                                          
author  = {{Omi Health}},                                                                                                 
year    = {2024},                                                                                                         
url     = {https://omi.health/research/omi-sum},                                                                          
note    = {3B parameter clinical SOAP model, fine-tuned from Phi-3, MIT licence}                                          
}                                                                                                                           
```                                                                                                                         
                                                                                                                          
## Ethical Considerations                                 

Users are urged to consider the ethical implications of AI in healthcare and ensure that any deployment of such models      
prioritizes patient safety and data privacy.
                                                                                                                          
## Contact                                                

For more information or to request access to Sum Small API, please contact [hello@omi.health](mailto:hello@omi.health) or   
visit [omi.health](https://omi.health)